Summary of Changes

New Files

    include/hbgtwide.h – public header for wide-character width calculation

Renamed / Refactored

    src/codepage/mk_wcwidth.c → src/codepage/hb_wcwidth.c (re-implemented with Harbour types, no macros)

Updated Makefiles & References

    src/codepage/Makefile – source list changed from mk_wcwidth.c to hb_wcwidth.c
    src/rtl/hbgtcore.c – include path switched to hbgtwide.h
    src/rtl/Makefile – adjusted -I path (../codepage → ../include)

Review Comments Addressed
 Source-level types instead of macros
    Uses Harbour types: HB_WCHAR32, HB_SIZE, HB_WCHAR.

 Header placed under include/ and renamed to hbgtwide.h

    Public include now matches project convention.

 HB_ namespace for all exported symbols

    mk_wcwidth() → hb_wcwidth()
    Helper tables/functions already prefixed (hb_wcswidth, hb_wcswidth_cjk).
    All widths returned as literal 0 / 1 / 2; no #define used.

 Character tables remain inside CODEPAGE modules

    combining[] and wide[] static tables live in src/codepage/hb_wcwidth.c.

Deferred for Future Work
Emoji support
Static Unicode TR-11 rules are used; emoji widths can vary by font—dynamic hooks can be added later.

Dynamic width tables
Maintainer noted potential need for runtime-replaceable tables; current code keeps static arrays but leaves the function-pointer path open.

Full glyph knowledge
As reviewer stated, “perfect timing” for complete glyph handling is not now; this patch provides the minimal, working foundation.

Merge branch 'fix-utf8-cursor-col' of github.com:woodhead2019/harbour into fix-utf8-cursor-col
Signed-off-by: woodhead2019 <woodhead2019@users.noreply.github.com>

Author: woodhead2019

src/codepage/mk_wcwidth.h include/hbgtwide.hsrc/codepage/mk_wcwidth.h renamed to include/hbgtwide.h

@@ -1,60 +1,31 @@
 /*
- * mk_wcwidth.h
+ * hbgtwide.h - Graphic Terminal Wide Character Width Support
  *
  * Copyright (C) 2001 Markus Kuhn <http://www.cl.cam.ac.uk/~mgk25/>
  *
  * This software is placed in the public domain.
  *
- * Historical reference: https://www.postgresql.org/message-id/attachment/8417/pg_mb_utf8.c
- *
  * Original source: http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c
  *
  * Adapted for Harbour by Dongming Wang <wangdongming / at / gmail.com>
  *
  * ANSI C89 compatible version for Harbour
+ *
+ * Unicode TR11 (East Asian Width) implementation
  */
 
-#ifndef MK_WCWIDTH_H
-#define MK_WCWIDTH_H
+#ifndef HBGTWIDE_H_
+#define HBGTWIDE_H_
 
-/* Include Harbour type definitions */
 #include "hbdefs.h"
 
-/* Type aliases: preserve original API */
-#define wchar_t  HB_WCHAR
-#define size_t   HB_SIZE
-
-/* Combining character width */
-#define COMBINING_WIDTH 0
-
-/* Zero width character */
-#define ZERO_WIDTH 0
-
-/* Full width character */
-#define FULL_WIDTH 2
-
-/* Half width character */
-#define HALF_WIDTH 1
-
-/* Wide character width */
-#define WIDE_WIDTH 2
-
-/* Narrow character width */
-#define NARROW_WIDTH 1
-
-/* Neutral character width */
-#define NEUTRAL_WIDTH 1
-
-/* Ambiguous character width */
-#define AMBIGUOUS_WIDTH 1
+HB_EXTERN_BEGIN
 
 /*
- * Function: mk_wcwidth
- * --------------------
  * Returns the width in screen columns of a Unicode code point.
  *
  * Parameters:
- *   ucs: Unicode code point (wchar_t)
+ *   ucs: Unicode code point (HB_WCHAR32)
  *
  * Returns:
  *   0: Control characters, non-printing characters, combining characters
@@ -67,15 +38,13 @@
  *   - Private use area characters are treated as narrow (width 1)
  *   - Unassigned characters are treated as narrow (width 1)
  */
-int mk_wcwidth( wchar_t ucs );
+HB_EXTERN int hb_wcwidth( HB_WCHAR32 ucs );
 
 /*
- * Function: mk_wcswidth
- * ---------------------
  * Returns the width in screen columns of a null-terminated Unicode string.
  *
  * Parameters:
- *   pwcs: Pointer to wide character string
+ *   pwcs: Pointer to wide character string (HB_WCHAR*)
  *
  * Returns:
  *   Total width of the string in screen columns
@@ -85,15 +54,13 @@ int mk_wcwidth( wchar_t ucs );
  *   - This function processes the entire string until null terminator
  *   - Returns -1 if any character has width 0 (non-printable)
  */
-int mk_wcswidth( const wchar_t *pwcs );
+HB_EXTERN int hb_wcswidth( const HB_WCHAR *pwcs );
 
 /*
- * Function: mk_wcswidth_cjk
- * -------------------------
  * Returns the width in screen columns of a substring of a Unicode string.
  *
  * Parameters:
- *   pwcs: Pointer to wide character string
+ *   pwcs: Pointer to wide character string (HB_WCHAR*)
  *   n: Maximum number of characters to process
  *
  * Returns:
@@ -104,6 +71,8 @@ int mk_wcswidth( const wchar_t *pwcs );
  *   - This function processes at most n characters
  *   - Returns -1 if any character has width 0 (non-printable)
  */
-int mk_wcswidth_cjk( const wchar_t *pwcs, size_t n );
+HB_EXTERN int hb_wcswidth_cjk( const HB_WCHAR *pwcs, HB_SIZE n );
+
+HB_EXTERN_END
 
-#endif /* MK_WCWIDTH_H */
+#endif /* HBGTWIDE_H_ */

src/codepage/Makefile

@@ -5,7 +5,7 @@ C_SOURCES := \
    cp_big5.c \
    cp_gbk.c \
    cp_utf8.c \
-   mk_wcwidth.c \
+   hb_wcwidth.c \
    cp_u16le.c \
    cpbg866.c \
    cpbgiso.c \

src/codepage/mk_wcwidth.c src/codepage/hb_wcwidth.csrc/codepage/mk_wcwidth.c renamed to src/codepage/hb_wcwidth.c

@@ -1,12 +1,10 @@
 /*
- * mk_wcwidth.c
+ * hb_wcwidth.c - Graphic Terminal Wide Character Width Implementation
  *
  * Copyright (C) 2001 Markus Kuhn <http://www.cl.cam.ac.uk/~mgk25/>
  *
  * This software is placed in the public domain.
  *
- * Historical reference: https://www.postgresql.org/message-id/attachment/8417/pg_mb_utf8.c
- *
  * Original source: http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c
  *
  * Adapted for Harbour by Dongming Wang <wangdongming / at / gmail.com>
@@ -15,20 +13,19 @@
  *
  * This is a simplified implementation of Unicode TR11 (East Asian Width).
  * It covers the most common Unicode character ranges.
- *
  */
 
-#include "mk_wcwidth.h"
-#include <stddef.h>
+#include "hbgtwide.h"
 
+/* Interval structure for binary search in range tables */
 struct interval
 {
    unsigned int first;
    unsigned int last;
 };
 
 /* Binary search in range table */
-static int bisearch( wchar_t ucs, const struct interval *table, int max )
+static int bisearch( HB_WCHAR32 ucs, const struct interval *table, int max )
 {
    int min = 0;
    int mid;
@@ -113,12 +110,10 @@ static const struct interval wide[] =
 };
 
 /*
- * Function: mk_wcwidth
- * --------------------
  * Returns the width in screen columns of a Unicode code point.
  *
  * Parameters:
- *   ucs: Unicode code point (wchar_t)
+ *   ucs: Unicode code point (HB_WCHAR32)
  *
  * Returns:
  *   0: Control characters, non-printing characters, combining characters
@@ -131,7 +126,7 @@ static const struct interval wide[] =
  *   - Private use area characters are treated as narrow (width 1)
  *   - Unassigned characters are treated as narrow (width 1)
  */
-int mk_wcwidth( wchar_t ucs )
+int hb_wcwidth( HB_WCHAR32 ucs )
 {
    /* Test for 8-bit control characters */
    if( ucs == 0 )
@@ -144,8 +139,6 @@ int mk_wcwidth( wchar_t ucs )
    if( bisearch( ucs, combining, sizeof( combining ) / sizeof( struct interval ) - 1 ) )
       return 0;
 
-   /* If we arrive here, ucs is not a combining or C0/C1 control character */
-
    /* Binary search in table of wide characters */
    if( bisearch( ucs, wide, sizeof( wide ) / sizeof( struct interval ) - 1 ) )
       return 2;
@@ -154,12 +147,10 @@ int mk_wcwidth( wchar_t ucs )
 }
 
 /*
- * Function: mk_wcswidth
- * ---------------------
  * Returns the width in screen columns of a null-terminated Unicode string.
  *
  * Parameters:
- *   pwcs: Pointer to wide character string
+ *   pwcs: Pointer to wide character string (HB_WCHAR*)
  *
  * Returns:
  *   Total width of the string in screen columns
@@ -169,13 +160,13 @@ int mk_wcwidth( wchar_t ucs )
  *   - This function processes the entire string until null terminator
  *   - Returns -1 if any character has width 0 (non-printable)
  */
-int mk_wcswidth( const wchar_t *pwcs )
+int hb_wcswidth( const HB_WCHAR *pwcs )
 {
    int width = 0;
 
    while( *pwcs != L'\0' )
    {
-      int w = mk_wcwidth( *pwcs );
+      int w = hb_wcwidth( (HB_WCHAR32)*pwcs );
 
       if( w < 0 )
          return -1;
@@ -188,12 +179,10 @@ int mk_wcswidth( const wchar_t *pwcs )
 }
 
 /*
- * Function: mk_wcswidth_cjk
- * -------------------------
  * Returns the width in screen columns of a substring of a Unicode string.
  *
  * Parameters:
- *   pwcs: Pointer to wide character string
+ *   pwcs: Pointer to wide character string (HB_WCHAR*)
  *   n: Maximum number of characters to process
  *
  * Returns:
@@ -204,13 +193,13 @@ int mk_wcswidth( const wchar_t *pwcs )
  *   - This function processes at most n characters
  *   - Returns -1 if any character has width 0 (non-printable)
  */
-int mk_wcswidth_cjk( const wchar_t *pwcs, size_t n )
+int hb_wcswidth_cjk( const HB_WCHAR *pwcs, HB_SIZE n )
 {
    int width = 0;
 
    while( n-- > 0 && *pwcs != L'\0' )
    {
-      int w = mk_wcwidth( *pwcs );
+      int w = hb_wcwidth( (HB_WCHAR32)*pwcs );
 
       if( w < 0 )
          return -1;

src/rtl/hbgtcore.c

@@ -58,7 +58,7 @@
 #include "hbapifs.h"
 #include "hbapierr.h"
 #include "hbapicdp.h"
-#include "../codepage/mk_wcwidth.h"
+#include "hbgtwide.h"
 #include "hbdate.h"
 #include "hbset.h"
 #include "hbvm.h"
@@ -92,7 +92,7 @@ void hb_gt_BaseUnlock( PHB_GT pGT )
  * This helper function checks the fWideCharWidth flag and returns
  * the appropriate display width for the character.
  *
- * When fWideCharWidth is enabled, uses mk_wcwidth() for accurate
+ * When fWideCharWidth is enabled, uses hb_wcwidth() for accurate
  * Unicode TR11 East Asian Width calculation. Otherwise returns 1
  * for backward compatibility.
  */
@@ -101,7 +101,7 @@ static int hb_gt_charDispWidth( PHB_GT pGT, HB_WCHAR wc )
    if( pGT->fWideCharWidth )
    {
       /* Use accurate Unicode TR11 width calculation */
-      return mk_wcwidth( (wchar_t)wc );
+      return hb_wcwidth( (HB_WCHAR32)wc );
    }
    /* Default: narrow character (width 1) for backward compatibility */
    return 1;